The Thunderstorms of 8 July 1989 in
the Northern Great Plains
Alexander H. Paul
Department of Geography
University of Regina
Regina, Sask. S4S OA2
and
Dan E. Blair
Department of Geography
University of Winnipeg
Winnipeg, Man. R3B 2E9
SUMMARY
This paper analyzes in detail an outbreak of severe thunderstorms in southeastern
Saskatchewan, southwestern Manitoba and North Dakota. Unexpectedly severe wind
damage, extensive crop-hail losses and very rapid translatory motion characterized these
storms. Data and information have been derived from official meteorological sources such
as weather stations, cooperative observers and regular pUblications including the Monthly
Record and Storm Data; and from unofficial sources such as local newspapers, crop-hail
insurance records and private individuals. Multi-million-dollar damages were incurred
even in this largely rural region, and the nature of the devastation at some localities was
suggestive of tornado occurrences. Despite the advent of weather radar, an effective
weather-watcher network throughout the study area, and an adequate job of forecasting
on both sides of the Canada-U.S. border, many local residents were unaware of the
significant threat which they faced. The paper concludes with some observations on the
problems at the general public's end of the weather warning issuance-disserninationreceipt-protection sequence.
INTRODUCTION
Drought conditions in 1989 over the eastern Canadian prairies and North
Dakota were eased in some localities by rainfall from scattered thunderstorms on
July 7-8. The severe storms of July 8, however, also did a great deal of crop and
property damage as they produced large hail and strong winds. Despite muchimproved forecasting of severe thunderstorms in the region in the 1980s these
storms still caught many people by surprise, and it is instructive to look at this
particular outbreak in more detail.
Figure I shows aspects of five separate named severe
thunderstorms in eastern Saskatchewan, western Manitoba and northern North
Dakota on Saturday, 8 July 1989 plus several other short-lived cells. The map is
based on a variety of official and unofficial information: AES and NOAA
precipitation records; crop-hail insurance data from Manitoba Crop Insurance
Corporation (MCIC), Co-operative Hail Insurance Company and Saskatchewan
Municipal Hail Insurance Association (SMHIA); reports from NOAA co-
170
Climatological Bulletin / Bulletin Climatologique 27(3), 1993
II •• d d·",.oc
--t
H., I
11 •••• J. w
-
I
--
A,t'·,
Kno."
160h.
.... - __ __
?I
\<ijoo •
d.~.a.
to
orop~rl1
2000
erop-h.11
Precipitation
"MOOSE JAil"
•c.u ••
"PEEBLES",
~
I
MANITOBA
SASKATCHEWAN
T
co .
MONTANA
•Wlillth"
• A, t
".
....
"
-
~
NORTH OAKOTA
- .J
" '-
...... -. _ _
,.,..t ... "
--_.
'Iif
C••
'-
so
l.
2100
d."'.ge
tree.
operative observer in North Dakota; coverage in daily and weekly newspapers
consulted at the Manitoba and Saskatchewan Provincial Archives; and other
miscellaneous sources.
The storms produced costly wind and hail damages at some
locations. Trailer homes were smashed in the Montmartre-Glenavon-Peebles area
of Saskatchewan' MCrC paid out more than 2.6 million dollars for hail damage;
according to NOAA's Storm Data, in North Dakota numerou buildings lost
roofs, and millions of dollars of crop-bail damage was done. Fortunately there
were only a few injuries and no fatalities. The troubling thing about the whole
episode is that the storms were more than adequately forecast and yet many
people still found lhemselves in dangerous situations. Perhaps this was because
over much of the region conditions were sunny, warm and dryas they had been
for much of the previous several weeks. The storms "seemed to come out of
nowhere", as a woman from the Roblin area of Manitoba put it; she was caught
outside in heavy hail and received "several hardball-sized bruises to show for it"
(Roblin Review, 11 July 1989).
Unexpectedly severe wind damage, extensive crop-hail losses and
very rapid translatory motion characterized these storms. Multi-million dollar
damages were incurred even in this largely rural region, and the nature of the
devastation at some places was suggestive of tornado occurrences.
THE INDIVIDUAL STORMS
Figure 1 shows the precipitation areas produced by the five principal storms of
July 8. Comparison with Figure 2, the map of daily rainfall listed for July 8 at
AES observing stations, is most instructive. The Moose Jaw storm is absolutely
invisible on Figure 2, a point which illustrates the well-known problem of
attempting to study thunderstorm precipitation from official records alone. The
Binscarth storm occurred in the early morning hours of July 8, and thus it would
show up only on the map of precipitation for the climatological day of July 7
(most co-operative observing stations take their daily reading of precipitation
amount at 0800 local time).
The Binscarth storm appears to have originated around
Bredenbury, Saskatchewan about 0300 CST and to have travelled southeast at
70-80 km/ h at least as far as Brandon, Manitoba where the AES reported hail at
0530 CST. It produced walnut-sized hail southwest of Binscarth and in
numerous other localities, and substantial crop losses on both sides of the
provincial boundary, although no significant property damage appears to have
resulted.
Ten hours later around 1500-1530 CST the Moose Jaw and Peebles
storms broke out. The former hit Moose Jaw city and agricultural areas
eastwards towards Regina with golfball-sized hail and heavy damage, but
fortunately just missed CFB Moose Jaw where the annual ~ir show was taking
172
Climatological Bulletin / Bulletin Climatologique 27(3), 1993
o
/'loose Jaw
°Wlnnlpeg
49 N
&stevan 0
_ S~S~. _
~IONT~
-
- -
:
-
-
-
N-: DAK.- -
50 kn.'
FIGURE 2
Daily Precipitati~n (mm), 8 July 1989
place. The Peebles storm did the most damage to property, with manyfarmsteads
and the entire village of Peebles being "mangled", according to the Regina
Leader-Post. Trailer homes were rolled and shattered but miraculously only a
few minor injuries resulted. The word "twister" was mentioned by several
eyewitnesses and there were two descriptions of a funnel cloud. The spatial
distribution of the wind damage leads us to believe that the occurrence of three
Fl or F2 tornadoes would be needed to explain the situation. In our view the
occurrence of one F2 tornado plus dO\vnburst or "microburst" winds (Fujita
1981) is the most likely explanation. Both these storms moved almost due east at
80 km/ h and both died out after only an hour or so.
About 1600-1630 CST, as the Moose Jaw and Peebles storms faded
away, the Roblin storm (Figure 1) developed south of Kamsack, Saskatchewan.
Most of its crop damage was done in Manitoba and it battered buildings and
cars in Roblin with hailstones up to 4 cm across. It continued east-southeast at
70 km/h and disappeared southeast of Ochre River around 2000 CST. At about
1630-1700 some short-lived but intense cells produced crop-hail losses southeast
of Regina and near Melville, and apparent downburst wind damage in and
around the village of Colfax, Saskatchewan (Figure 1).
The final major storm of July 8, the North Dakota storm, was also
the largest and longest-lived. It was more or less contemporary with the Roblin
storm, tracking in the same direction at the same speed, but 240 km to the south.
It developed in the vicinity of Alameda, Saskatchewan, laid a swath of damaging
hail at least 20 km wide in the extreme southeast of the province and caught the
A.H. Paul & D.E. Blair / Thunderstorms of 8 July 1989
173
10';)
FIGURE 3 Surface weather analysis, OOOOZ 9 J].Ily 1989 (18 CST 8 July)
courtesy AES
very southwest corner of Manitoba from Pierson south. Wind damage along the
southern edge of the hailswath in North Dakota was spectacular. Storm Data
reported a tornado at Sherwood, North Dakota where the roof was ripped off a
school gymnasium. NOAA co-operative observers reported many trees downed
in the Clark Salyer National Wildlife Refuge and severe damage in the town of
Maddock. Kramer, Esmond and Tunbridge had much wind damage to solidly
built houses and 20,000 acres of crops were reported destroyed in Benson
County. Hailstones 2.5 inches (6.3 cm) in diameter fell in Newburg and Kramer
(Figure 1). The North Dakota storm persisted for at least 4 hours and had a
path length of more than 300 km.
THE WEATHER SITUATION
At the synoptic scale, a weak surface low-pressure centre moved across southern
Saskatchewan on July 8 (Figure 3). On the evening of July 7 and through the
night, strong thunderstorms developed in the over-running warm moist air over
southwestern Manitoba, with some just extending into southeastern
Saskatchewan. Satellite imagery indicates at least one mesoscale convective
174
Climatological Bulletin / Bulletin Climatologique 27(3), 1993
complex (Maddox, 1980) with high-level cold anvil clouds coalescing for a
distance of about 300 krn over southwestern Manitoba in the early hours of July
8. A westsouthwest 500 mb flow with a core of maximum speed over western
Montana curved anticyclonically to flow about due west to east over
southeastern Saskatchewan at 0600 CST.
Surface winds during morning and afternoon of July 8 over the
region were light to moderate southeasterly; thus thunderstorms would develop in
a strongly sheared environment. Dewpoints were high in the eastern portion of
the warm sector close to the warm front, hitting 18-19° C at many stations by
noon local time (Regina had its highest dewpoint of the year, 20.4° C, at 1700
CST). Convective instability was present and some thunderstorms were
experienced northeast of the warm front, for instance at Brandon and Dauphin
during the morning. Severe weather watches and warnings were in effect in
southwestern Manitoba at this time. Satellite imagery shows that conditions were
essentially clear, however, over the regions that were affected later in the day.
Even the photos for 1401 CST show virtually no convection in the warm sector
of the surface low.
The 1431 CST satellite photo shows an isolated narrow band of
towering cumulus aligned WSW-ENE, centered over Regina and about 100 km
long and 10-15 km wide. The "cold cloud shield" (Maddox 1980) from this line
of convection is clearly visible on the 1501 CST photo and the line has moved
ESE. Between 1400 and 1500 CST at Regina the surface wind shifted from E19
to NNE30, dewpoint rose from 17.8° C to 18.8°C and dry-bulb temperature
remained at 29° C. The beginnings of the Moose Jaw storm were now visible
from the satellite along the southern edge of an area of broken clouds marking
the cold front. On the 1531 CST photo the Moose Jaw storm is easily identifiable
as a separate entity from the much larger cloud shield which has developed from
the Peebles storm, but by 1601 CST it appeared to be merging with this cloud
shield which now covered a larger portion of southeastern Saskatchewan. An
extension towards the southeast of this cloud shield indicates that the North
Dakota storm has now reached severe stage; on the 1701 CST photo a strong
new cell was developing right along the 49th Parallel and the cloud shield over
southeastern Saskatchewan and southwestern Manitoba has the dimensions of a
mesoscale convective complex.
Within the warm sector there had been a marked dryline all
through the day, and at 1800 CST the surface analysis suggests that the dryline
met the centre of the surface low in the vicinity of the intersection of the
Manitoba-Saskatchewan-N. Dakota borders. At this time the North Dakota
storm was located right at this same position. The satellite photos suggest that
the Peebles storm had also developed along this dryline; the precise origin of the
Moose Jaw storm is more questionable.
A.H. Paul & DE. Blair / Thunderstorms of 8 July 1989
175
DISCUSSION
A better meteorological understanding of storms of this type has emerged in
recent years, but this case study of 8 July 1989 strongly suggests that such
understanding is not being transmitted to the pUblic. In Saskatchewan the local
and regional newspapers almost all described the wind damage as being due to
tornadoes. At Maddock, North Dakota, however, the damage was of the same
intensity as at Sherwood where Storm Data reported a tornado. At Maddock a
wind gauge hit 78-82 mph (l30 km/h) but no one described the damage as being
due to anything more than very strong winds. This observed windspeed of 130
km/h, with a speed of storm translation of 75 km/h, could be caused by a
downburst producing an outflow at the surface of 55 km/h relative to the storm.
A photograph of the gust front of the North Dakota storm taken
with a telephoto lens just prior to the storm's arrival at Maddock at about 1940
CST was provided to us by Mr. Fred Rehling, the local co-operative observer for
NOAA. The gust front has given rise to a spectacular roll cloud when, the warm
air ahead of the storm is lifted above the outflowing downdraught. Although
there is some indication of rotating motion in the roll cloud, this cloud appears to
have a horizontal extent of several kilometres; at least at this stage of the storm's
life there is no tornado funnel in the picture. Yet the storm "took two grain
elevators [and] roofs off many homes, etc." in the town. It appears to be a
microburst situation.
AES personnel on the prairies tend to be conservative in using the
word "tornado" in post-storm damage evaluations where no very definite
observations of funnel clouds touching ground were made. With the severe
weather watch program and extended weather radar coverage in the 1980s, there
is greater awareness by prairie meteorologists of the frequency of severe
thunderstorm events and forecasters also have been rather more prepared to issue
tornado watches since the Edmonton disaster of 31 July 1987. However, the
problem of getting the message across to the public about damaging
thunderstorm winds still remains. The storms of 8 July 1989 occurred during a
severe thunderstorm watch but this watch was not upgraded to a tornado watch.
The result seems to have been that the public was not aware of the threat which
they faced on this date. One might speculate that the public feels that if
tornadoes or severe wind damages are going to occur, then a tornado watch will
always be issued. They do not realize that microbursts can be just as damaging
as - and more widespread than - an Fl or even an F2 tornado. The usual
terminology of "damaging wind gusts" and the "possibility of tornadoes"
accompanying severe thunderstorms seems to be inadequate to convey to the
public the real potential for injuries and even fatalities which these storms carry
even when they do not generate true tornadoes.
There are certainly some misconceptions in many people's minds
about damage potential from thunderstorm winds. An excellent example of this
176
Climatological Bulletin / Bulletin Climatologique 27(3), 1993
is furnished by the coverage in the Weyburn Review newspaper of the
windstorms on 8 July 1989. The damage in and near Colfax was quite well
reported, but the comment was then made that the "twister" was not experienced
in the villages of Fillmore, Osage and Lang, 40 km, 30 km and 30 km
respectively in various directions from Colfax. We are not blessed with
clairvoyance but this does seem to us to suggest that if the reporter visualized a
"twister" as being a small tornado, then he/ she had little understanding of the
characteristics of the dimensions and paths of tornadoes.
A further comment in this same newspaper article is very revealing.
On 8 July 1989 there was some wind damage in the city of Weyburn itself. A
large tree was uprooted, a sure indication of gusts exceeding 70 km/ h, and
shingles blown off a number of roofs. The reporter commented that winds at
Weyburn that afternoon averaged only 40 km/h, a "mild breeze" compared to
the winds at Peebles, but that this "mild breeze" had still been enough to
dislodge shingles and uproot the large tree and some "shrubbery". The reporting
seems to indicate a lack of awareness that (a) damaging winds in one part of the
city may not affect the location of the weather station 1 or 2 km away; (b)
reported average windspeeds say nothing of short-period maximum gusts that
may have occurred; and (c) the average of 40 km/h is insufficient to explain the
damage done.
In conclusion, we believe that despite all the excellent initiatives
that have been taken by the AES in the forecasting and warning of severe
thunderstorm weather in the prairie region, a communications problem with the
general public still exists. This case study of 8 July 1989 indicates that while
many people reacted very competently to the storm situation, there were many
others in the region who were unaware that they faced a significant threat, who
have little awareness of the nature of thunderstorm wind damages, and who are
thus unable to interpret the true meaning of forecasts and watches even if they
hear them. Whether the AES with its limited resources can do much more in this
context is questionable. Perhaps the real need is for the provinces, municipalities
and emergency-preparedness organizations to undertake at the local level a
program of raising awareness of the thunderstorm wind hazard on the prairies.
ACKNOWLEDGEMENTS
Thanks are due to the various organizations and individuals mentioned in the
paper for permitting access to their data and information. We are particularly
grateful to Ed Lozowski for making the University of Alberta archive of satellite
imagery available to us, and to numerous helpful personnel of AES in
Winnipeg. Financial support was received from SSHRC and the University of
Regina. Thanks to Donna Glass for word processing and to David Ackerman
for Figure 1. An earlier version of this paper was presented at the 26th Annual
CMOS Congress in Quebec City in June 1992.
A.H. Paul & D.E. Blair / Thunderstorms of 8 July 1989
177
REFERENCES
Fujita, T.T., 1981. Tornadoes and downbursts in the context of generalized planetary
scales. Journal afthe Atmospheric Sciences, 38: 1511-1534.
Maddox, R.A., 1980. Mesoscale convective complexes. Bulletin of the American
Meteorological Society, 61: 1374-1387.
NOAA 1990, Storm Data, Vol. 31, No. 7, July 1989, National Oceanic and Atmospheric
Administration, National Climatic Data Center, Asheville, North Carolina.
LOCAL NEWSPAPERS:
Kipling (Sask.) Citizen, 17 July 1989.
Moose Jaw (Sask.) Times-Herald, 10 July 1989.
Roblin (Man.) Review, 11 July 1989.
Weyburn (Sask.) Review, 12 July 1989.
178
Climatological Bulletin
I Bulletin Climatologique 27(3), 1993